Seminar Report Extensibility and Data Sharing in Evolving
Seminar Report Extensibility and Data Sharing in Evolving Multi-Tenant Database Reporter: Qiu Lei
About Author • Stefan Aulbach • Technische Universität München. Institut für Informatik • Research Area • • • Multi-Tenancy Databases and Software-as-a-Service Next-Generation Databases Auto. Globe: An Adaptive Infrastructure for Distributed Services • Publication List • • Extensibility and Data Sharing in Evolving Multi-Tenant Databases(ICDE 2011) A Comparison of Flexible Schemas for Software as a Service(SIGMOD 2009) Multi-Tenant Databases for Software as a Service(SIGMOD 2008) Ruminations on Multi-Tenant Databases(BTW 2007)
Background • Essential features of enterprise applications • • • Data Sharing (master data should be shared rather than replicated for each tenant) Extensibility (application modification and extension, which applies both to the database schema and master data it contains) Evolution (evolution of the schema and master data, which occurs as the application and its extensions are upgraded) • The typical Saa. S environment • the latest available version of the applicaion is the most popular and a limited number of tenants lag one or more versions behind • Optimized for main-memory • allows to identify which versions of the instances are currently used by a given set of tenants, and which can be removed from main-memory
Contributions • FLEXSCHEME • a data management model for Multi-Tenant DBMSs • XOR Delta • • a physical data organization approach of segment sequences (there is a definition of segment sequences in FLEXSCHEME)
Requirement • Saa. S applcations develop in at least two dimensions: Extensibility and Evolution
Data Management Model • The global view with three Tenants(Fig. 3) • The local view of Tenant T 3(Fig. 4)
FLEXSCHEME • The relation across all tenants • (e. g. T 1, T 2, T 3)
FLEXSCHEME • The relation for a given tenant • (e. g. T 3)
Physical Data Organization Of Segment Sequences • Snapshot Approach • Data Structures • Pointwise Access • Scan Access • Creation of New Segments • Purge Operation • Space Efficiency
Physical Data Organization Of Segment Sequences • Dictionary Approach • Data Structures • Pointwise Access • Scan Access • Creation of New Segments • Purge Operation • Space Efficiency
Physical Data Organization Of Segment Sequences • Temporal Approach • Data Structures • Pointwise Access • Scan Access • Creation of New Segments • Purge Operation • Space Efficiency
Physical Data Organization Of Segment Sequences • Differential Delta Approach • Data Structures • Pointwise Access • Scan Access • Creation of New Segments • Purge Operation • Space Efficiency
Physical Data Organization Of Segment Sequences • XOR Delta Approach • Data Structures • Pointwise Access • Scan Access • Creation of New Segments • Purge Operation • Space Efficiency
XOR Delta Approach • Pointwise Access
Evaluation • Data Access
Evaluation • Space Requirements
Evaluation • Maintenance Operations
Seminar Report Thank You!
- Slides: 18